A Convenient Procedure for the Quantitative Conversion of Reactive

Philip Dimitrov , Jack Emert , Jun Hua , Sandor Keki , and Rudolf Faust. Macromolecules 2011 44 (7), 1831-1840. Abstract | Full Text HTML | PDF | PDF ...
1 downloads 0 Views 516KB Size
1090

BROWNAND REI

VOL. 31

A Convenient Procedure for the Quantitative Conversion of Reactive Alcohols and Olefins into the Corresponding Chlorides HERBERT C. BROWNAND MIN-HONREI Richard B. Wetherill Laboratory, Purdue University, Lafayelte, Indiana Received December 2, 1966 ‘The hydrogenator previously described, which generates hydrogen automatically from sodium borohydride to achieve the quantitative hydrogenation of unsaturated compounds, has been adapted to the hydrochlorination of reactive alcohols and olefins. I n this procedure hydrogen chloride is generated automatically as required to react with the alcohol or olefin, and the generation ceases when the reaction is complete. In this manner it is possible both to follow the rate of utilization of the hydrogen chloride and to convert the alcohol or olefin essentially quantitatively into product without excessive contact with the hydrochlorination reagent.

Reactive alcohols, such as t-butyl alcohol, are readily converted into tertiary chlorides by being shaken at room temperature with excess concentrated hydrochloric acid.1s2 Alternatively, anhydrous hydrogen chloride may be passed into the alcohol until saturation has been achieved. I n the course of our investigations dealing with the nature of the intermediate formed in the solvolysis of bicyclic derivatives, 4,5 we encountered difficulties in applying these simple procedures to relatively labile bicyclic structures, Excess treatment of the alcohols with the reagent led to the formation of appreciable quantities of undesired by-products. On the other hand, inadequate treatment led to the formation of mixtures of the alcohol and chloride which proved exceedingly difficult to separate. Consequently, we were led to explore the possibility of developing a simple procedure which would permit us to follow the uptake of hydrogen chloride quantitatively, so that the reaction coulcl be halted as soon as all of the alcohol had been converted into the chloride, thereby avoiding either the incomplete conversion or the too exhaustive treatment that had plagued our attempts to utilize the older procedures. Recently, a simplej but highly effective, apparatus was described for the automatic generation of hydrogen as required for the hydrogenation of unsaturated corn pound^.^^^ It appeared that the same apparatus and principle might be adapted to the hydrochlorination of resictive alcohols and olefins. Accordingly, we undertook to study this possibility. Indeed, we discovered that ue could convert many of the bicyclic alcohols and certain reactive olefins essentially quantitatively into the corresponding chlorides at 0’ in a reaction time of approximately 100 sec. Consequently, the procedure :?ot, only solved our immediate problem, but it opened up the possibility of reexamining the hydrochlorination of many reactive terpenes where the precise nature of the initial hydrochlorination product may be :.n doubt.

Procedure We utilized the commercial model of the preparative hydrogenator.7 However, in order to avoid possible (1) J. F. Norris and A. W. Olmeted, “Organic Syntheses,” Coll. Vol. I, John Wiley and Sons, Inc., New York, N. Y . , 1941, p 144. (2) H. C. Brown and R. S. Fletcher, J . A m . Chem. Soc., 71, 1845 (1949). (3) H. C. Brc.wn and R. B. Kornblum, ibid., 76, 4510 (1954). (4) H. C. Brown and F. J. Chloupek, i b i d . , 81, 2322 (1963). (5) H. C. Brown, F. J. Chloupek, and M.-H. Rei, ibid., 86, 1246 (1964). (6) C. A. Brown and H. C. Brown, ibid., 84, 2829 (1962). (7) Commercial models of these automatic hydrogenators, both preparative and analytical, are currently available from the Delmar Scientific Laboratories, In&, Maywood, 111. 60154.

corrosion of the hvpodermic needle of the buret in the preparative unit; we substituted the glass-tipped buret and the corresponding valve of the analytical unit for these two items of the preparative unit. In this way the concentrated hydrochloric acid in the buret comes in contact only with glass and the mercury in the valve before it reacts with the concentrated sulfuric acid in the generator to produce anhydrous hydrogen chloride. The complete assembly is shown in Figure 1. The following procedure will convert from 10 to 100 mmoles of alcohol or olefin into the chloride. However, it can readily be modified to handle either smaller or larger quantities. Approximately 200 ml of concentrated sulfuric acid is placed in a 500-ml modified erlenmeyer flask, serving as the generator G. A Teflon-coated magnetic stirring bar is introduced. (It is important that the sulfuric acid be efficiently stirred to effect thorough mixing of the less dense hydrochloric acid phase with the sulfuric acid.) The slight curvature of the flask bottom and the Teflon collar on the stirring bar are not essential for the present application, but they greatly facilitate smooth, efficient stirring of the contents. The apparatus is assembled so that it is gas tight. In the buret B is placed an adequate amount of concentrated hydrochloric acid. The height of the liquid column must be controlled so that it is supported by the column of mercury in the valve V. By maens of a small rubber bulb attached to the top of the buret, pressure was supplied to force 3 to 4 ml of concentrated hydrochloric acid through the mercury seal to generate sufficient hydrogen chloride to flush the system. (The amount was selected to provide a volume of hydrogen chloride that was approximately twice the free volume of the system.) The reaction flask R, usually a 125-ml modified flask, for the usual scale of the preparations, was commonly immersed in an ice bath to provide a convenient reaction temperature. However, in many cases it proved quite satisfactory to carry the reaction out a t room temperature, and in some cases quite low temperatures were used. The reaction is initiated by injecting, by means of a hypodermic syringe, through the introductory port P, 10 to 100 mmoles of the alcohol or olefin, neat, or as a concentrated solution in n-pentane, methylene chloride, or carbon tetrachloride. Reaction begins immediately. (It is important that the contents of the reaction flask be stirred vigorously by the usual

APRIL1966

CONVERSION OF REACTIVE ALCOHOLS AND OLEFINSINTO

Teflon-coated bar and magnetic stirrer.) As soon as the pressure drops 10 to 20 mm below atmospheric, concentrated hydrochloric acid is drawn through the mercury seal in the valve V and it then reacts with the magnetically stirred sulfuric acid to generate hydrogen chloride to restore the pressure. The reaction proceeds automatically until absorption of hydrogen chloride ceases, as shown by the behavior of the mercury in the bubbler, or the constancy of the quantity of hydrochloric acid remaining in the buret. Experiments revealed that under the operating conditions each milliliter of the concentrated hydrochloric acid generated 11.0 f 0.1 mmoles of hydrogen chloride. I n the case of olefins, such as a-methylstyrene, 1 mole of hydrogen chloride was taken up for each mole of olefin. However, in the case of the alcohols, additional hydrogen chloride was utilized to saturate the water produced in the reaction. At 25' each mole of alcohol utilized 1.29 moles of hydrogen chloride, at 0' the corresponding figure was 1.35 moles. In the case of liquid chlorides, prepared neat from the corresponding alcohols, it was adequate to remove with the aid of a capillary tube the small lower phase of water saturated with hydrogen chloride. After a brief aspiration to remove traces of dissolved hydrogen chloride, the organic phase appeared to be essentially pure tertiary chloride by all of the usual tests. I n cases where solvents were employed, it was removed under vacuum to recover the chloride, or the solution was cooled to low temperatures to precipitate the solid chloride.

Figure 1.-Apparatus

100

90

(8) To test the utility of the crude chlorides thus prepared, without intermediate purification for rate measurements, we solvolysed the products in 80% aqueous ethanol directly and observed the following first-order rate constants measured at : ! 5 O 0 031 hr-1 for t-butyl Chloride, 0 055 hr-1 for t-amyl chloride, 0 098 l i r - 1 for triethylcarbinyl chlonde. Evidently, chlorides prepared in this manner can be utilized directly for rate measurements

CHLORIDES

1091

used for the automatic hydrochlorination of alcohols and olefins.

-

70 60 40 80

Results and Discussion The three aliphatic alcohols, t-butyl alcohol, tamyl alcohol, and triethylcarbinol were each injected neat into the reaction flask, maintaining a slow constant rate of stirring in the reaction flask. Under these conditions the reaction required 75, 15, and 7.5 min, respectively, to proceed to completion. The relative magnitudes of the reaction times correspond to the relative ease of formation of the three carbonium ions, as measured by the rates of solvolysis at 25' in 80% aqueous ethanol: 0.033, 0.055, and 0.099 hr-l, respectively.2z8 The time required for reaction is a function of the efficiency of stirring in the reaction flask. Thus, by using more vigorous agitation, the time required to convert t-amyl alcohol to its chloride was reduced from 15 to 1.3 min. It is probable that the apparatus could be developed into a satisfactory procedure for determining the rate of reaction of such alcohols with hydrogen chloride, but this will require vigorous agitation of the contents of the reaction flask adequate to avoid diffusion i-ontrol of the observed rate. Since this TTas not a part of our current objective, we did not attempt to develop such a procedure. Perhaps the most unexpected result of ous study was the observation that many of the more reactive alcohols, including the bicyclic alcohols of immediate

THE

B .* -c

50

P;

30

-

10 20

I

0

I

100

200

I

I

300

Time, sec.

Figure 2.-Hydrochlorination

of various alcohols and olefins a t 0'.

concern to our studies, underwent conversion to the tertiary chlorides with remarkable ease, the reaction being essentially complete in 2 min at 0'. Since we are doubtless operating in the range where diffusion control must be important, this means that the observed reaction time, 2 min, must represent a maximum, with the actual reaction being considerably faster than that observed. Representative data are shown graphically in Figure 2. (The interesting behavior of 2-methyl-endo-norborneol will be discussed later.) In a number of cases we compared the relative merits of n-pentane, methylene chloride, and carbon tetrachloride as solvents for the reaction. While the results

1092

BROWN AND REI

VOL. 31

TABLE I DATAFOR

Alcohol

THE

mmoles

HYDROCHLORINATION OF REPRESENTATIVE TERTIARY ALCOHOLS

Solvent

Temp, "C

Time.' min

Hydrogen chloride used, mmoles

Alkyl chloride obtained, mmoles

r

T+D

or

mp, 'C

Product Puritg by titrn

Yield isold, %

42.1 Neat Roomtemp 75 38.3 31 99 74 39.1 Neat Roomtemp 15 37.3 38.2 1,4058 100 97 40.0 Neat Roomtemp 7 . 5 39.9 37.4 1.4330 100 94 10.0 CHzClz 0 2.5 10.3 9.3 1.5957 95 93 10.0 CH2Clz 0 3.5 9.5 9.0 (60). (90)~ 18.0 n-CSHiz 10 3.5 17.5 17.5 1.4460 100 96 n-CJh 20.1 10 26 20.3 17.6 20-25 98 88 20,O n-CSHlz 0 12 20.3 19.4 120-122 99 97 20.1 n-CgH12 0 6.5 21 .o 18.7 43-45 99 94 10.0 n-CsHi~ 0 1.8" 9.9 10.0 n-C5H12 0 2.25 9.6 10.0 CHzClz 0 1.2d 10.1 10.0 CHzClz 0 1.Od 10.5 The reaction time is a function of the stirring rate. The values reported should only he considered to be an indication of the very fast rates realized. * No detectable vinyl proton and hydroxy group was detected in nmr and infrared spectra. c The tertiary chloride is highly unstable and the product appears to lose hydrogen chloride during the aspiration procedure. d The data indicate a faster rate in methylene chloride t4hanin n-pentane.

t-But,yl alcohol t-Pentyl alcohol Triethylcarbinol Benzhydrol Diphenylmethylcarbir 01 1-Methylcyclopentanol 2-Methyl-endo-norborneol 1,2-Dimethyl-endo-norborneol 2-Phenyl-endo -norborneol

(1

100

-

90

-

80

-

70

,-

40 30

2o 10 0

t

1, v 0

200

400

600 800 1000 1200 1400 Time, see.

Figure 3.---Rate of absorption of hydrogen chloride by amethylstyrene a t 0" : neat ( X ), in methylene chloride (2 M , A), and in n-pentane solution ( 2 M , 0).

hint that methylene chloride is preferable for the hydrochlorination reaction itself, the advantage is very minor. On the other hand, n-pentane possesses advantages for the recovery of product by removing the solvent under vacuum, as well as for crystallizing out a solid product a t low temperatures. Finally, the use of carbon tetrachloride possesses obvious advantages where it is desired to subject the initial product directly to nmr examination without isolation. We also explored the effect of temperature. We were able to convert t-amyl alcohol to its chloride in a reasonable time at 25, 0, and - 4 5 O . Even the secondary alcohol, benzhydrol, absorbed the theoretical quantity of hydrogen chloride in 2.5 min a t 0' in methylene chloride solution. Consequently, we tended to carry out our preparations a t Oo, except in cases, such as 2-methyl-endo-norborneol, where the 0' reaction proved to be relatively slow.

As is shown in Figure 2, all of the compounds absorbed the theoretical amount of hydrogen chloride within 2-3 min, vith one exception, 2-methyl-endonorborneol. This alcohol rapidly absorbed 65% of the calculated quantity of hydrogen chloride a t Oo, and exhibited no further observable change in the following 10 min. Then a slow further uptake of hydrogen chloride became evident, approaching the theoretical amount over a relatively long period. Similar behavior was observed at 10'. In this case the second stage was considerably faster, the reaction proceeding to essential completion in 26 min. The initial plateau corresponds closely to the rapid formation of a l : l addition compound, which is then transformed only relatively slowly into water, which takes up the additional hydrogen chloride, and the tertiary chloride. The 1:1 addition compound is probably the oxonium salt, ROH2+C1-. The difference in the behavior of the ezo and endo isomers presumably arises from the same factors which are responsible for the difference in the rates of solvolysis of oxo- and endo-norbornyl derivative^.^ Typical data for the hydrochlorination of representative tertiary alcohols are summarized in Table I. Finally, we examined the hydrochlorination of amethylstyrene as a representative olefin. At 0' the hydrochlorination proceeded rapidly and quantitatively, being complete in approxiniately 2 min (Figure 2). At -45' the reaction proceeded equally rapidly. However, a t room temperature the addition of hydrogen chloride to the olefin was much slower.'O Similar negative temperature coefficients for the addition of hydrogen chloride to norbornenell and isobutylene12 have been described.

.

(9) For a discussion with pertinent references, see H. C. Brown, F. J ChIoupek, and M.-H. Rei, J. Am. Chem. Soc., 86, 1248 (1964). (10) For the fast reactions we are obviously operating within the rate of diffusion control. Consequently, we cannot conclude from the data at 0 and -45' that the true reaction rates are the same. However, it is safe to conclude from the data that the reaction rate a t room temperature is much lower than the rates a t 0 and -45'. (11) L. Schmerling, J . Am. Chem. Soc., 68, 195 (1946). (12) F. R. Mayo and J. J. Kata, ibid., 69, 1339 (1947).

CONVERSION OF REACTIVE ALCOHOLS AND OLEFINSINTO

APRIL1966

I n the hydrochlorination of a-methylstyrene, the use of methylene chloride as a solvent appears to have significant advantages over n-pentane in favoring the rate of reaction (Figure 3). A number of runs was carried out with a-methylstyrene to test the reproducibility of the procedure, as well as to test the effect of solvent on the reaction rate. Representative data are summarized in Table 11. TABLE I1 HYDROCHLORINATION OF WMETHYLSTYRENE AT 0"

DATAO N a-Methylstyrene, mmoles

THE

Solvent

Time, sec

Hydrogen chloride used, mmoles

Molar ratio HCl/olefin

39.2 0.98 40. la Seat 126 138 10.9 1.03 10.5 Neat 10.1 1.01 126 10.0 Xeat 1.06 12s 10.0 9.4 Keat 1.01 130 10.0 9.9 Neat 9.6 0.98 9.8 Seat 136 1.01 995 10.1 10.0 n-CjH1:b 1.03 10.3 10.0 n-CjHi:b 1100 1.06 61 10.6 CHQClp' 10.0 1.05 10.5 10.0 CHQC12' 123 10.3 1.03 CH2C12' 120 10.0 a Isolated a 97.5% yield of t-cumyl chloride, 99.5% pure by titration, ~ * O D 1.5230. 2 M solution.

Conclusions

It is evident from these results that the automatic hydrogenator previously reported can be readily adapted for the automatic hydrochlorination of reactive alcohols and olefins. This opens up the possibility of reexamining the hydrochlorination of many terpenes and other labile olefins where a knowledge of the structure of the initial reaction product would be of interest. We are currently exploring this possibility. Finally, it; appears evident that the general principal appears applicable to many reactions in which one wishes to achieve the automatic, quantitative addition of a gas to a liquid or solid phase. Experimental Section Stoichiometry.-The commercial hydrochloric acid utilized was 12.1 M . However, hydrogen chloride possesses a modest solubility in sulfuric acid containing several per cent of water. Consequently, it was necessary to establish precisely how much hydrogen chloride would be delivered by the addition of 1.00 ml of the concentrited hydrochloric acid to the sulfuric acid in the generator.

THE

CHLORIDES

1093

This was done in two ways. One milliliter of water was injected by means of a hypodermic syringe into the reaction flask of the apparatus and the generation of hydrogen chloride was allowed to proceed to completion. The water saturated with hydrogen chloride was titrated, and the value was compared with the number of milliliters of concentrated hydrochloric acid delivered from the buret. Alternatively, successive portions of a-methylstyrene were injected into the reaction flask and the amount of concentrated hydrochloric acid delivered by the buret was noted. Both procedures gave a figure of 11.0 f 0.1 mmoles of hydrogen chloride generated from each milliliter of the concentrated hydrochloric acid delivered into the generator. Similarly, by determining the amount of hydrogen chloride required to saturate 1.00 ml of water a t 0 and 25', the figure was arrived a t that each mole of tertiary alcohol would require 1.35 moles of hydrogen chloride at 0" and 1.29 moles of hydrogen chloride a t 25" to achieve the conversion to the chloride and the saturation of the aqueous phase produced in the reaction. These figures correspond to the values actually observed for the reaction of 2-phenyl-endo-norborneol in the apparatus. Hydrochlorination of 2-Methyl-endo-norborneol.-The apparatus was assembled and flushed with hydrogen chloride as is described in the Procedure. The reaction flask was immersed in a bath a t lo", and the reaction was initiated by injecting 20 mmoles (2.53 g) of 2-methyl-endo-norborneol in 10 ml of npentane solution through the injection port. There was an immediate uptake of approximately 75% of the required amount of reagent. Then the reaction proceeded slowly to 1OOyoutilization, being essentially complete in 26 min. The organic layer %-as withdrawn from the reaction flask and subjected to gentle aspiration to remove dissolved hydrogen chloride. Fresh pentane was added to bring the volume to precisely 10.0 ml, and a 1-ml aliquot was added to 80yc aqueous ethanol a t room temperature. Standard alkali was added to neutralize the hydrogen chloride produced, the amount corresponding to 97.4% of reactive chloride in solution. This quantity is listed as the purity (titration) in Table I. The remaining solution was evaporated under vacuum to constant weight. There was obtained 2.54 g,lS 17.5 mmoles, of 2-methyl-ezo-norbornyl chloride, mp 2O-25".I4 A weighed sample of the chloride was solvolyzed in 80% aqueous ethanol at 25'. There was produced 98.6% of the calculated quantity of hydrogen chloride. Hydrochlorination of a-Methylstyrene .-In this case the reaction flask was cooled to 0" and the reaction was initiated by injecting 40 mmoles (4.73 g) of a-methylstyrene, ~ W D1.5381. After 10 min, the product was subjected to gentle aspiration to remove dissolved hydrogen chloride. There was obtained 6.02 g (39.2 mmoles) of 2-chloro-2-phenylpropane, n% 1.5230. A sample of this chloride was solvolyzed in SOYo aqueous ethanol. The hydrogen chloride produced was between 99 and 100% of the theoretical.

Acknowledgment.-This investigation was assisted by Grant 19878 provided by the National Science Foundation. (13) Corrected for the aliquot removed in the determination of the yield by titration. N. J. Toivonen, E. Siltanen, (14) The pure chloride possesses m p 25': and K. Ojala. Ann. Acad. Sci. Fennicae, Ser. A II,64 (1955).